12 x 4110 MOSFET Extreme Modder Controller LYEN Edition $129

[Zenid]

I bought the 12-FET version of this controller about a week ago, and with the exception of one minor gripe I've been very happy with it so far.

On my initial test drive, though, with the controller set to 50A/125A rated/phase. I noticed something odd just now and again: I found that this was if I tried to start from a dead stop - or very slow - with the throttle wide open. There would be a 'thunk' as power to the motor would cut off. It wouldn't pull power unless I opened the throttle more gradually, then fully once I was moving.

However this seemed to go away after the first journey. Yesterday, though I raised the current to 55A/135A, and the problem came back with a vengeance - clearly there is some kind of current limitation in operation here. Lyen said the controller was capable of pulling 65A continuous in its existing form, but clearly things aren't set up right or another modification needs to be made to enable this to happen.

I'm hearing lots of talk about "shunt values", and the fact that they sometimes have to be changed to enable you to draw more current. Am I going to have to make a change here too?

If so, then this doesn't sit easy with Lyen's assurance that the I can draw 65A from this controller in its current form, as that clearly does not appear to be the case...

[Lyen]

Hi Zenid,

I have sent you a PM to clarify your concerns.

Regards,
Lyen

I bought the 12-FET versiion of this controller about a week ago, and with the exception of one minor gripe I've been very happy with it so far.

On my initial test drive, though, with the controller set to 50A/125A rated/phase. I noticed something odd just now and again: I found that this was if I tried to start from a dead stop - or very slow - with the throttle wide open. There would be a 'thunk' as power to the motor would cut off. It wouldn't pull power unless I opened the throttle more gradually, then fully once I was moving.

However this seemed to go away after the first journey. Yesterday, though I raised the current to 55A/135A, and the problem came back with a vengeance - clearly there is some kind of current limitation in operation here. Lyen said the controller was capable of pulling 65A continuous in its existing form, but clearly things aren't set up right or another modification needs to be made to enable this to happen.

I'm hearing lots of talk about "shunt values", and the fact that they sometimes have to be changed to enable you to draw more current. Am I going to have to make a change here too?

If so, then this doesn't sit easy with Lyen's assurance that the I can draw 65A from this controller in its current form, as that clearly does not appear to be the case...

[icecube57]

Im going to clear a few things up for Lyen. I did some research on your scooter and there are a few details left out. Im not being a dooche but Im just helping you understand whats going on. The scooter has an unladen weight of 104kg. Before a rider is put on and before extra batteries are even put in the scooter. The 2 extra 12v 28ah batteries weigh 40lbs a piece. So you have an upgraded vehicle weight of 308lbs before you put a rider on. We are going to say you are an average 180lbs. so thats a total of 488lbs you are asking the motor to move at any given time. Thats a tall order. Im a big bloke myself at 315lbs but my bike weighs 50lbs with batts. So for me thats only 365lbs. The motor is a tiny casted mag wheel motor motor thats similar to 50cc gas scooters. The specs say that the scooter motor is rated for 1400w stock. One thing about small motors like this is that they have a very low turn and high inductance. Meaning they draw alot more current and are wound to go faster to make up for the small wheel size. So you have a "LEAD SLED" and undersized controller for the job an and wondering why you are hitting the current limit of the controller. I would have strongly advise you to upgrade to an 18fet right away. It seems like you need alot of current to feed you hungry motor and the 12fet isnt up to the job in its stock condition it has to be really beefed up. It seems like the performance you are after and the motor you have... you are going to need 75-100A to get the going for a dead stop without it hitting the current limit of the controller. One thing about these low turn motors is they arent very forgiving. They only think of themselves in operation. They want all the current they can take and dont care if the controller can keep pace with it or not. Everything else will die before the motor will. Ive ran a low turn count motor and its a blast when fed properly but its a pain in the ass and its a costly road to walk to find that butterzone to where you and the motor is happy.


Oh yeah.... keep playing around and bumping the current limit up you will surely kill your controller. I personally dont go above 45-125 myself. One thing that may save you from purchasing a new controller is Block Time in the Parameter Settings Program. Maybe increasing it to 3 seconds will allow the controller to dump enough current into the motor to get you off the line. Block Time is basically a window of a few seconds to where the motor is allowed to draw whatever current it needs to get going... during this time the controller is sorta dazed and its severely raped in the process.

It snaps out of it starts regulating the current down near the programmed value. Due to the characteristics of you motor and its selfish hoarding mentality its advised you leave your BlockTime at 1 but if you are risky... bump it up to 3 you "should" be ok. If you are stupid take it to 5 or higher and see what happens. I bet you will be back on here so fast blaming everyone else for your misfortune. =P I gave you warning.

[aaronski]

my Lyen 12 fet is set to 65/65, with no issues. when I set the phase currents that high I also see it cutting out to protect the controller if I floor it from a dead stop. I actually have it on my "to Do" list to try every phase setting at 65 amp battery settings and see where the limits are. With any luck I'll have that info on Saturday to give everyone a feel for where this controller freaks out.

also, thanks for the Block time info. if it's unlimited, won't it trip just like it does when phase amps are set too high?

[icecube57]

Ive been meaning to try out equal phase and battery current. I think you are missing out a little on torque by setting your phase that low. You will be suprise at the increase. But stock.... 120A phase is about high as I will go and maybe 50-60A on the battery.

The glorious thing about Block Time is that it basically shuts off the protection temporarily during hard launch/acceleration for a certain amount of seconds that you program. But this can be dangerous because you can blow mosfets doing this paired with the wrong motor. I would play around with higher values on a 5304 5305 or a 6x10 or 5x12 9C only because the resistance and High BEMF of those motors quickly choke the amps down but I wouldnt do it on a 5302 or a 5303 or a 12x5 or 10x6 9C you rick blowing your controller because they will request high phase current all the way to top speed because of their low resistance and the lack of the ability to generate sufficent BEMF....with low turn count the controller almost sees them as a dead short and they are VERY hard on controllers.

The controller can peak for a short time or in burst but it cant sustain it.

[Zenid]

I would have strongly advise you to upgrade to an 18fet right away. It seems like you need alot of current to feed you hungry motor and the 12fet isnt up to the job in its stock condition it has to be really beefed up. It seems like the performance you are after and the motor you have... you are going to need 75-100A to get the going for a dead stop without it hitting the current limit of the controller. One thing about these low turn motors is they arent very forgiving. They only think of themselves in operation. They want all the current they can take and dont care if the controller can keep pace with it or not. Everything else will die before the motor will. Ive ran a low turn count motor and its a blast when fed properly but its a pain in the ass and its a costly road to walk to find that butterzone to where you and the motor is happy.
{...} Oh yeah.... keep playing around and bumping the current limit up you will surely kill your controller. I personally dont go above 45-125 myself.

Thanks for the feedback and the useful info on block times. However the reason I got the version of this controller, and that Lyen recommends it is because it is well suited to this type of bike, and is the exact type that many other owners have upgraded to. I'm well aware of the issues of current limitation, which is why I opted for this IRFB4110 12-FET one by Lyen, who said initially that this is capable in its 'factory form' of pulling 65A continuous. He's since revised this to more like 60-65A, but with modifications (reinforcing the tracks and so forth) it's capable of handling up to 100A at 100V, according to the blurb.

The Ego Scooter and other lookalikes can run fine on as little as just 35A, but most of us aren't happy with the sluggish acceleration and low speeds so make various mods like this to make them better. Lyen's made some helpful suggestions about why I'm getting this dickiness with the throttle on higher settings, and why it stutters or triggers the cut-out, so I'll look into them. But it is starting to look like I might well have to modify the shunt as others have done to 'trick' the board into giving more power if I want to push the motor any harder. I'm just interested in exploring the limits of the performance of my bike, and identifying the weak links in the system.

The main weak-link that we've identified is with the phase-wires, which are rather thin on the 1400W motors many of us have and tend to get hot when we ramp up the current, - we therefore keep a close eye on these as we experiment with uprated settings/mods on controllers. I was interested with what you were saying about the motors themselves being capable of handling whatever we can throw at them, as that would seem to support what many of us are considering doing, - that is cracking open the hub motor and upgrading the phase wires to a higher capacity AWG rating.

No, I'm not going to come back and start blaming people for a wrecked controller. I'm very careful with testing and modding these things, and do so with extreme caution and only after getting a range of opinions and advice.

And don't worry. I don't think it's 'being a douche' at all for you to offer your opinion. I welcome the feedback, and any input and new info is always appreciated, thanks!

[Zenid]

my Lyen 12 fet is set to 65/65, with no issues. when I set the phase currents that high I also see it cutting out to protect the controller if I floor it from a dead stop. I actually have it on my "to Do" list to try every phase setting at 65 amp battery settings and see where the limits are. With any luck I'll have that info on Saturday to give everyone a feel for where this controller freaks out.

also, thanks for the Block time info. if it's unlimited, won't it trip just like it does when phase amps are set too high?

You're using 65A for both rated AND phase current? Everybody else I know is using a phase current of around 2.5 times the rated current (such as 30/75A, 50/125A etc.). How do we decide what multiple to use, and with what kind of motor? Is it just trial and error, or is there some reasoning behind this multiple. Nobody I've asked seems to understand why this figure (2.5) is used and what can happen if others are used instead, except that there can be 'odd effects'...

[aaronski]

You're using 65A for both rated AND phase current? Everybody else I know is using a phase current of around 2.5 times the rated current (such as 30/75A, 50/125A etc.). How do we decide what multiple to use, and with what kind of motor? Is it just trial and error, or is there some reasoning behind this multiple. Nobody I've asked seems to understand why this figure (2.5) is used and what can happen if others are used instead, except that there can be 'odd effects'...

Yes, currently 65/65 as a quick fix for the cutout issue on my 26" BMC v2 motor. and yes, the low end torque is awful untill around 5mph. I really need to fix it.
It cut my startup watts from 5k down to 2k, rising to three at around 10mph, then dropping back down again.

I should have some hard numbers when I get some time to fiddle with it later this weekend. Or perhaps I'll ride to work tomorrow/friday at 95/125 and compare them all.

[icekreme1002]

I have recently been discussing this ratio with Lyen for my 18fet/9c combo. I am still in newb territory with all this but from my understanding from 2.3-2.6 ratio makes sense.
Here is Lyens explanation to me via email....
"The way to explain the ratio in plain English is pulse-width modulation current output to the motor is not the same as the battery current. It is kind of a voltage converter thing that convert from 12v to 24v as an example. But it inverts in current, not voltage. Although you can set it to 50/150 or more but the difference is very minor because the the controller need to ask for certain amount of current from the battery in order to give X amount of maximum current output to the motor. You can try 50/150 or 180 and see what I mean. Another example I can think of in plain English is you try to borrow money to buy a $10000 car from the bank. But the bank requires you to put at least $3000 down payment in order for them lend you $7000."

So basically you can set phase to a really high number, but if the controller cant get what it needs from the battery setting it higher wont make a difference.

The 2.3-2.6 ratio represents the range where battery amps will be able to provide what the motor is demanding in phase amps.
Hopefully that helped some other new kids like me

[bmxed]

There would be a 'thunk' as power to the motor would cut off. It wouldn't pull power unless I opened the throttle more gradually, then fully once I was moving.

However this seemed to go away after the first journey. Yesterday, though I raised the current to 55A/135A, and the problem came back with a vengeance - clearly there is some kind of current limitation in operation here. Lyen said the controller was capable of pulling 65A continuous in its existing form, but clearly things aren't set up right or another modification needs to be made to enable this to happen.

Hi Zenid,
This is probably no fault of the Lyen controller. See if you have a BMS in your battery pack and what the current rating is. It might be sensing an over-current condition and is shutting down to prevent damage to the battery.

[bigbore]

In my location e-bikes must go lower 25km/h (15,5mph) when assisted by the motor and the rider must pedal if he wants to be assisted by the motor or in other words the motor must assist the rider only if he is pedaling and must stop the assistance at the speed of 25km/h.

I would like to know from you if is possible with your controller to build a dual mode e-bike


I need a standard mode running within laws in case you are catched by the cops and a "turbo" mode with a speed of 55-65 Km/h. The standard mode must be switched very quickly and not to be reversed in case you are got by the cops while running a "little fast". I want a jumper that I can pull and hide in my trouser pocket so without the jumper the bike runs at 15mph and the motor is powered only if the pedals are spinning.
How does the controller recognize the spinning pedals? Do I need a sensor?

Please let me know if it's possible.
Regards, Simone.

[dodjob]

Hi bigbore,
The only solution I know is to use a Cycle Analyst and set it to 25Km/h. Then you install a switch on the override channel. This means that when the switch is close the CA will override the throttle signal and cut the acceleration at 25Km/h. When the switch is open, nobody interfer with you and the controller you can drive as fast as you want
There is a thread somewhere If you search a little you'll find it
Gruß,
H.

*edit* yeah and the pedal sensor has to remain active...

[Zenid]

Hi Zenid,
This is probably no fault of the Lyen controller. See if you have a BMS in your battery pack and what the current rating is. It might be sensing an over-current condition and is shutting down to prevent damage to the battery.

Hi bmxed. I am not running off a lithium pack yet. My bank is an SLA one, so there is no BMS acting as intermediary. It is therefore the controller cutting off power - most probably due to the way it is set up. Lyen has been very helpful in suggesting some things I could change in the settings, like specifying 120 degree hall sensors.

Here's a picture of my 'engine' BTW. I have 170A welding cable connecting my battery bank so clearly there's no bottle neck there...:



It is also possible that the faster drain is hard for my SLAs to keep up with, and that the voltage drop as I draw current under high load is simply triggering the lvc, so I can try tinkering with that too (though I am surprised that the lvc would response to very transient 'dips' of). Another possibility is that the high current spike is triggering the controller's safety cut-off.

Either way, it's just a case of tinkering with it until I find the right combination of settings. I'm otherwise perfectly happy with the controller, which has greatly improved the power I can get out of my bike. I just like to tinker to see if I can push the envelope just a little further...

What I'm looking at now, is the trick of reprogramming the board as a 218 instead of a 212, to 'fool' the controller into pulling more power. Am I right in thinking that I would have to raise my current settings by a third to compensate for this, or do I LOWER them by a third. I have seen contradictary information on this point, and don't want to make a mistake here and blow up my FETs

What does everyone think? Is it just a really bad idea to do this without modifying the shunt or reinforcing the tracks? And how do I modify the shunt in the way people are talking about to raise the current it can draw? Lyen gave me a link, but it spoke of modding power-side resistors, and all I have on my board are shunts like this:
http://zenid10.files.wordpress.com/2011/03/p1020162.jpg

What's the difference between these two methods (software versus hardware) of upping the current limit, and how is this accomplished on this board? Any help and advice here greatly appreciated. I'm not about chance anything unless I'm absolutely sure that I know what I'm doing...

[skeetab5780]

i can probably guarantee this is not your issue, but even if your cable is rated at 170Amps it looks like your still using 30Amp rated crimp-on style ring/fork terminals on all of your phase wires. It looks nice and clean, but they're rated for a reason.

hook a Voltage meter up to your SLA bank while ur riding it, and within 1 minute you'll know if its voltage sag causing the problem or not. Process of elimination FTW.

if its not the Voltage sag the next bet is hitting the Max Amp setting on the controller, from a stand-still moving all that weight
can spike a ridiculous amount of Amperage.

I have the Lyen 72v 30A controller and have not had an issue like this..but i also ride a converted mountain bike and im not that heavy (180lb)

Good luck trouble-shooting your issue! keep us posted on the solution.

[icecube57]

You go down in controller size to bump up the current. Your controller needs the power traces reinforced before you can do something like this. It can boost you current by almost double. I had a reinforced 12 fet running a 5303 at 66v and 45A and it still got hot enough to melt the solder on the board. I still suggest you get an 18fet. I doubt you pulling enough current to cause enough sag to trip the lvc although the symptoms you list do suggest that. We really need voltage under load of your pack and then on a per battery before we can rule that out. I still think its a low turn motor that just sucking alot of current. JOHNINCR can give you some insight on this. The only thing you can Really do is go big or go home.

This isnt going to end well for you. You insist on pushing the boundary limits of the controller.... Even with my 5303 which is a very hard motor to drive I didnt have overcurrent issues with it cutting out but it did blow mosfets even on an 18fet. You keep bumping the current and trying to get more amps out of a stock unit without beefing it up it will end in a FAIL. Controllers are expensive and in your case LYEN is not local enough for you to send it to him for repair.

[amberwolf]

I need a standard mode running within laws in case you are catched by the cops and a "turbo" mode with a speed of 55-65 Km/h. The standard mode must be switched very quickly and not to be reversed in case you are got by the cops while running a "little fast". I want a jumper that I can pull and hide in my trouser pocket so without the jumper the bike runs at 15mph and the motor is powered only if the pedals are spinning.
How does the controller recognize the spinning pedals? Do I need a sensor?

You'd need the pedalec sensor for the latter, if the controller has a separate pedalec input. Otherwise you'd need a pedal-based throttle.

For the former, you'd need a "jumper" that when installed turns *off* the limits, by a set of relays or similar. When the jumper is installed, the relays disconnect the pedalec loop and the speed limiter loop on the controller itself, if it has them.

If you do not want to disable the pedalec part and intend to pedal all the time anyway (recommended to keep people from noticing you as much ) then you only need a little plastic card like they use on treadmills, which would hold apart a little leaf switch or a roller on a microswitch (recommended for weatherproofing). That switch would go to the speed limiter jumper on the controller, enabling the limit when the card is pulled.

The card could be anything, including an ID card on a lanyard around your neck, or a piece of jewelry on a long necklace, that would pull out of the slot automatically as you dismount or even when you lean back far enough.

[Zenid]

This isnt going to end well for you. You insist on pushing the boundary limits of the controller....

Did you miss the bit where I said "I'm not about chance anything unless I'm absolutely sure that I know what I'm doing..."?

I'm not "insisting" at all. I'm simply trying to understand the limits of this controller by getting input from folk like you who clearly have a better understanding of the issues than I do. I'm very happy with the performance of it at this level, but I've seen others who have made modifications and I'm simply exploring the viability of getting my system to cope with more power. I GREATLY appreciate your advice and warning, but please rest assurred that I'm not dismissing your 18-FET advice, it's just that this is what Lyen advised me to get, this is what I bought, and this is what I'm stuck with now.

It's perfectly good and I'm very happy with the added power and speed, I just heard that a lot of folk have done clever things (like brutally reinforce the tracks with copper wire etc) to pull a bit more power out.

Even with my 5303 which is a very hard motor to drive I didnt have overcurrent issues with it cutting out but it did blow mosfets even on an 18fet. You keep bumping the current and trying to get more amps out of a stock unit without beefing it up it will end in a FAIL. Controllers are expensive and in your case LYEN is not local enough for you to send it to him for repair.

You're preaching to the converted! I fully understand that traces need beefing up and other modifications need to be done to get better performance, and painfully aware of the consequences of destroying a perfectly good controller! I'm asking these questions so I can get the benefit of your wisdom and understand the options I have and the risks they carry. The fact that the current limiter is being tripped (possibly) is obviously a warning that I should heed, like you say. I just thought perhaps I could push things a little further without undue risk of damage providing I monitor temperatures and keep them within safe boundaries. England is chilly, - we have no shortage of cold air here.

Your stern warnings are very much appreciated and well-taken! Thanks

You keep mentioning motor types and numbers like "5303", and say that the low turn count of my motor is probably guzzling up current and liable to melt things if I mess with the current limitation. Where would I look to identify the type of motor I'm using (without actually opening the hub motor). Is there something on the case, perhaps?

Thanks again for your input, and please relax, as I'm in no hurry to do anything stupid!

Incidentally, I followed Lyen's advice and switched from the default option for auto-select and specified "120 degree", which is what my motor is. I'll see how I get on with this...

[Zenid]

i can probably guarantee this is not your issue, but even if your cable is rated at 170Amps it looks like your still using 30Amp rated crimp-on style ring/fork terminals on all of your phase wires. It looks nice and clean, but they're rated for a reason.

Very well observed! But are they rated at any less than the wires themselves? You're quite right in that this seems to be one clear, weak link in this system, and others with the same bike are talking about upping the AWG rating on the phase wires to the motor. I also admit to being unsure of what this implies in terms of the number. I was assuming that three phase wires in parallel with a 30A limit means 90A total limit. I'm just going on what other people with this bike have been doing...

Your note of warning here is well taken. One poor soul in our forum did this to his junction block with an 18-FET controller :

hook a Voltage meter up to your SLA bank while ur riding it, and within 1 minute you'll know if its voltage sag causing the problem or not. Process of elimination FTW.

That was my next plan But I have to set it up so I have a meter taped onto my instrument panel and wires run under the seat and tedious stuff like that, and haven't got round to it yet!

if its not the Voltage sag the next bet is hitting the Max Amp setting on the controller, from a stand-still moving all that weight can spike a ridiculous amount of Amperage.

I have the Lyen 72v 30A controller and have not had an issue like this..but i also ride a converted mountain bike and im not that heavy (180lb)

Good luck trouble-shooting your issue! keep us posted on the solution.

Thanks It's only been a real problem when I set the controller to above 50A, but it has been dicky on a couple of occasions at this level too, usually once the batteries have been run down a little or worked hard for a while. It might just be a combination of lvc and SLAs struggling to keep up. I've only noticed it when I've been harsh on the throttle for a while.

Then again it might be the controller limit... No matter, it isn't a big deal as mostly it runs fine. I did 28 miles on it yesterday (with a long recharge) and it only happened once.

My latest thing was to follow Lyen's advice and set the motor type to "120 degree" rather than "auto-select", as he also said it could be the hall sensor electronics getting confused if the wheel starts off at a certain angle of rotation (or something like that). I did that today, and so far so good. It's just a matter of tweaking and experimenting until it goes away. I plan to upgrade to LiFePO4 in the near future anyhow, and losing 30Kg should help things enormously!

[amberwolf]

I was assuming that three phase wires in parallel with a 30A limit means 90A total limit.

Unfortunatley, no, they're not actually in parallel. Instead, they're in series (or rather, two at a time are, with the third unused at that moment).

So 30A is all you get thru there, if that's the limit of the phase wires and/or connectors and or windings themselves.

[Zenid]

I was assuming that three phase wires in parallel with a 30A limit means 90A total limit.

Unfortunatley, no, they're not actually in parallel. Instead, they're in series (or rather, two at a time are, with the third unused at that moment).

So 30A is all you get thru there, if that's the limit of the phase wires and/or connectors and or windings themselves.

Well, hang on, from my limited understanding of how that PWM stuff works, doesn't it fire a very quick burst down each phase wire consecutively (or something like that) but always such that a single wire is off most of the time? You say 'two at a time' - is that right? Whatever it is, then surely you have to factor in the 'off' time as a running average...

So how do you factor that into the guage of the phase wires & connectors you should have for a particular system?

[amberwolf]

Well, hang on, from my limited understanding of how that PWM stuff works, doesn't it fire a very quick burst down each phase wire consecutively (or something like that) but always such that a single wire is off most of the time? You say 'two at a time' - is that right? Whatever it is, then surely you have to factor in the 'off' time as a running average...

Since it requires a complete power path to get the current thru the motor, one phase wire is "positive" while another is "negative" at any one time. It rotates thru them in sequence to get it thru there.

Off time for any phase wire is only 1/3 of the time, at the least. Can be more if it is PWMing to limit current or voltage, but during block commutation or at no limits max throttle, it's only 1/3 for a three-phase BLDC like a typical ebike motor.

There is a form of wiring and controller using a Wye termination that actually grounds the center wire and so has four "phase" leads to the controller, but I've never actually worked with a controller that does that. That would only pass current thru one phase wire at a time, plus the ground (center) wire. Off time in that case would be minimum of 2/3, except for the common wire, which would never be off except possilby during PWM off events (maybe not then, if the phases overlap right).

More typically, the phases are just tied together at one end for Wye, and that bundle is tied off, not connected to anything else. Thus to get current to flow, two phases are in operation at any one time.

So how do you factor that into the guage of the phase wires & connectors you should have for a particular system?

Same as if only one were in use, keeping in mind that you have twice the wire length involved at any one moment that you thought you did.

[Zenid]

Same as if only one were in use, keeping in mind that you have twice the wire length involved at any one moment that you thought you did.

Ah I see. But doesn't the third of the time that a single wire is in its 'off' state, effectively introduce a factor of two-thirds into the amperage rating that the phase to have to support the power? For example, a 50A rated setting would only require 33.3A rated phase wires, and so on?

[Zenid]

FYI, the second part of my blog/review of the LYEN 12-FET controller detailing my experiences with it is now available for the perusal of anyone who is interested:
http://zenid10.wordpress.com/2011/04/10/50-at-50-amps/
(Part one: http://zenid10.wordpress.com/2011/04/04 ... and-angry/)

[amberwolf]

Ah I see. But doesn't the third of the time that a single wire is in its 'off' state, effectively introduce a factor of two-thirds into the amperage rating that the phase to have to support the power? For example, a 50A rated setting would only require 33.3A rated phase wires, and so on?

Well, no, becuase if you want ot put a certain current thru a wire, it should be sized for the max current you're going to push thru it (it doesn't have to be, but it'll get hotter if it's not than if it is).

However, you can think of it that way if you like, and don't mind the extra heat in the wires--it may not be enough to make a difference depending on your setup, wire thickness, voltae, current, etc. If there is sufficient airflow to cool the wires it also may not matter, at least outside the motor (inside it doesnt' matter because it's so hot from the motor's own waste heat that heat from the short bit of phase wires is probably irrelevant).

[Zenid]

Well, no, becuase if you want ot put a certain current thru a wire, it should be sized for the max current you're going to push thru it (it doesn't have to be, but it'll get hotter if it's not than if it is)

How I understood PWM to work is that by firing lots of tiny pulses it 'fools' the motor into thinking it is getting higher and higher voltages, because the wiring effectively 'sees' just a continuous average voltage of the square wave-form it gets.

In other words if I put a pulse of 60A down a wire for 1ms, then switch it off for the next 1ms then give it another pulse at 60A for then next 1ms and so on, then the wire will respond to all intents and purposes as if it were a continuous 30A flow, and will therefore generate the same heat as if it just had 30A running through it. In a similar way, A/C 'tricks' the wire into believing it has 0 amps average (or something odd like that) by a similar arcane bit of physics/math trickery, if I recall from my Physics class.

If I'm wrong about this, please feel free to correct me. I don't want to "think of it like that" unless this also how the wire "thinks of it", as this makes all the difference between frying my wires, and not!

Perhaps someone else could also weigh in and offer their cents worth, as it's something that I really need to properly understand if I'm to see to it that my wiring is adequate to cope with any increased load I subject it to... How much amperage do those three wires actually 'see' if the controller is putting out, say, 50A?